We present a novel image-based method for separating diffuse and specular reflections of real objects under distant environmental illumination. By illuminating a scene with only four high frequency illumination patterns, the specular and diffuse reflections can be separated by computing the maximum and minimum observed pixel values. Furthermore, we show that our method can be extended to separate the diffuse and specular component under image-based environmental illumination. Applications range from image-based modeling of reflectance properties to improved normal and geometry acquisition.